Dynamic reconfiguration of resilient logical modules in a software defined server

1. A computer system, comprising:

a plurality of physically interconnected computing nodes, wherein a guest operating system is executing on a virtual machine comprising a set of virtualized resources provided by a distributed virtual machine manager comprising a plurality of hyper-kernels running on the plurality of physically interconnected computing nodes, and wherein an application is running on the guest operating system;

wherein an indication is received of a request to remove a first computing node from the plurality of physically interconnected computing nodes;

wherein, in response to receiving the indication of the request to remove the first computing node, and while the application is running, a first hyper-kernel running on the first computing node migrates virtualized resources associated with the first computing node from the first computing node to one or more other computing nodes in the plurality of physically interconnected computing nodes;

wherein migration of virtualized resources to the first computing node is blocked;

wherein upon completion of migration of virtualized resources from the first computing node, the first hyper-kernel transmits one or more messages to at least some of the other computing nodes in the plurality of physically interconnected computing nodes indicating that migration of virtualized resources from the first computing node is complete; and

wherein a second hyper-kernel on a second computing node in the plurality of physically interconnected computing nodes:

receives a message indicating that migration of virtualized resources from the first computing node is complete; and

determines locally whether a virtualized resource is still on the first computing node, wherein based at least in part on a local determination that the virtualized resource is still on the first computing node, the second hyper-kernel sends a location request for the virtualized resource.

2. The computer system of claim 1 , wherein migrating the virtualized resources associated with the first computing node comprises migrating, from the first computing node, at least one of virtualized processors, virtualized I/O devices, or virtualized physical memory.

3. The computer system of claim 2 , wherein migrating a virtual processor from the first computing node comprises invoking, by the first hyper-kernel, a stall.

4. The computer system of claim 1 , wherein migration of virtualized resources to the first computing node is blocked at least in part by transmitting one or more messages to at least some of the other computing nodes in the plurality of physically interconnected computing nodes indicating that the first computing node is no longer a destination for migration of virtualized resources.

5. The computer system of claim 1 , wherein in response to migrating of a virtualized resource from the first computing node to a second computing node, the first hyper-kernel transmits one or more location update message to at least some of the other computing nodes in the plurality of physically interconnected computing nodes indicating that the virtualized resource has moved from the first computing node to the second computing node.

6. The computer system of claim 1 , wherein the first hyper-kernel receives one or more responses from computing nodes in the plurality of physically interconnected computing nodes acknowledging receipt of at least some of the one or messages indicating that migration of the virtualized resources from the first computing node is complete.

7. The computer system of claim 6 , wherein based at least in part on receiving a response acknowledging receipt of a message indicating that migration of the virtualized resources from the first computing node is complete, the first computing node is transitioned out of an operational state.

8. A method, comprising:

receiving an indication of a request to remove a first computing node from a plurality of physically interconnected computing nodes, wherein a guest operating system is executing on a virtual machine comprising a set of virtualized resources provided by a distributed virtual machine manager comprising a plurality of hyper-kernels running on the plurality of physically interconnected computing nodes, and wherein an application is running on the guest operating system;

in response to receiving the indication of the request to remove the first computing node, and while the application is running, migrating, by a first hyper-kernel running on the first computing node, virtualized resources associated with the first computing node from the first computing node to one or more other computing nodes in the plurality of physically interconnected computing nodes;

blocking migration of virtualized resources to the first computing node; and

upon completion of migration of virtualized resources from the first computing node, transmitting, by the first hyper-kernel, one or more messages to at least some of the other computing nodes in the plurality of physically interconnected computing nodes indicating that migration of virtualized resources from the first computing node is complete;

wherein a second hyper-kernel on a second computing node in the plurality of physically interconnected computing nodes:

receives a message indicating that migration of virtualized resources from the first computing node is complete; and

determines locally whether a virtualized resource is still on the first computing node, wherein based at least in part on a local determination that the virtualized resource is still on the first computing node, the second hyper-kernel sends a location request for the virtualized resource.

9. The method of claim 8 , wherein migrating the virtualized resources associated with the first computing node comprises migrating, from the first computing node, at least one of virtualized processors, virtualized I/O devices, or virtualized physical memory.

10. The method of claim 9 , wherein migrating a virtual processor from the first computing node comprises invoking, by the first hyper-kernel, a stall.

11. The method of claim 8 , wherein migration of virtualized resources to the first computing node is blocked at least in part by transmitting one or more messages to at least some of the other computing nodes in the plurality of physically interconnected computing nodes indicating that the first computing node is no longer a destination for migration of virtualized resources.

12. The method of claim 8 , wherein in response to migrating of a virtualized resource from the first computing node to a second computing node, transmitting, by the first hyper-kernel, one or more location update message to at least some of the other computing nodes in the plurality of physically interconnected computing nodes indicating that the virtualized resource has moved from the first computing node to the second computing node.

13. The method of claim 8 , comprising receiving, by the first hyper-kernel, one or more responses from computing nodes in the plurality of physically interconnected computing nodes acknowledging receipt of at least some of the one or messages indicating that migration of the virtualized resources from the first computing node is complete.

14. The method of claim 13 , wherein based at least in part on receiving a response acknowledging receipt of a message indicating that migration of the virtualized resources from the first computing node is complete, the first computing node is transitioned out of an operational state.